A Preliminary Method for Tracking In-Season Grapevine Cluster Closure Using Image Segmentation and Image Thresholding

Figure 1

General workflow of the proposed method; steps 1 and 2 include image collection and annotating part of the collected image dataset; step 3 includes running the first image segmentation algorithm (PSPNet) to segment images into the cluster and noncluster class; step 4 includes multiple steps of image processing such as image masking, grayscale conversion, and image contrast enhancement (see methodology section); step 5 is second image segmentation where cluster areas were segmented into berry pixels and gap pixels using Otsu’s thresholding method; step 6 uses berry area and gap area to calculate % CC.